| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a USDA-ARS, Box 646402, Washington State Univ., Pullman, WA, 99164
b The Noble Foundation, Ardmore, OK, 73401
c Dep. of Statistics, Washington State Univ., Pullman, WA, 99164. Mention of product names does not represent and endorsement of any product or company but is given only to clarify the methodology; other products may be equally effective
* Corresponding author (rcjohnson{at}wsu.edu).
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
) usually correlates with the dry matter to transpiration ratio (transpiration efficiency) in C3 species, but its heritability and relationship to forage production are less clear. In a 4-yr study of tall fescue (Festuca arundinacea Schreb.) at Pullman, WA (relatively cool with low humidity), and Ardmore, OK (relatively hot with high humidity). we determined (i) if differences in divergently selected populations made on single plants were maintained in solid seeded plots, and (ii) how in selected populations and a set of four cultivars was related to forage production. Differences in for low and high populations selected on spaced plants were maintained in solid seeded plots at both Pullman and Ardmore. At Pullman, the low selection had higher production than the high selection with the base population intermediate. Partial correlations with all entries between and forage production, controlling for harvest date effects, were not significant. However, partial correlation between and forage production on the selected and base populations was significant (r = –0.59, P < 0.05, n = 12) at Pullman, although not at Ardmore. The data show selection for low may improve forage production in some environments, although not consistently. For breeding tall fescue, one cycle of phenotypic selection for low in advanced material is recommended.
Abbreviations: TE, transpiration efficiency
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
) estimates TE in C3 species at the whole plant level (Farquhar and Richards, 1984; Hubick et al., 1986; Read et al., 1991; Johnson and Bassett, 1991; Johnson and Tieszen, 1994; Johnson et al., 1995; and Impa et al., 2005). Leaf gas exchange studies have also shown that the negative correlation between TE and extends to fundamental photosynthetic processes, that is, the carbon assimilation rate to stomatal conductance ratio (Read et al., 1991; Johnson, 1993; Johnson et al., 1995).
Differences in result mostly from differences in fractionation of 12CO2 and 13CO2 by ribulose-1,5-bisphosphate carboxylase/oxygenase during photosynthesis (Farquhar et al., 1989). The physiological basis for the negative correlation between and TE in C3 plant species is that it estimates the tissue internal leaf substomatal CO2 to ambient CO2 concentration integrated over time. A lower internal CO2 to ambient CO2 concentration promotes higher TE and results in lower . Internal substomatal CO2 concentration is balanced between stomatal conductance to CO2 and the tissue capacity for carboxylation (Farquhar et al., 1989). The relationship between and TE, as derived by Farquhar et al. (1989), is linear and negative, so that lower indicates higher TE.
The expectation is that indirect selection for TE through would improve productivity of plants, especially under drought conditions. Research into the putative link between and productivity has ranged from no relationship (Johnson et al., 1995; Matus et al., 1995; Menéndez and Hall, 1996) to positive (Condon et al., 1987; Araus et al., 2003) to a negative relationship (Rebetzke et al., 2002). A positive -to-productivity relationship suggests that lower TE is advantageous, whereas a negative relationship suggests that higher TE is advantageous.
Studies have often, but not always, shown that the heritability of is sufficiently high that genetic gain in TE would be expected in a selection program. Broad-sense heritability for peanut (Arachis hypogaea L.) was estimated at 0.53 (Hubick et al., 1988) and for wheat (Triticum aestivum L.) at 0.61 (Ehdaie and Waines, 1994). Menéndez and Hall (1996) also reported intermediate values for broad-sense heritability (0.33–0.47) in cowpea [Vigna unguiculata (L.) Walp.], but realized heritability was only 0.06 to 0.19. Realized heritability values of bean (Phaseolus vulgaris L.) were low, ranging from 0 to 0.12 under irrigated and rainfed environments (White, 1993). Narrow-sense heritability values for greater than 0.75 were reported by Read et al. (1993) for crested wheatgrass [Agropyron desertorum (Fisch. ex Link) Schult.] and from 0.47 to 0.63 in three wheatgrass species (Frank et al., 1997).
Clearly, , and presumably TE, can be affected by genetic as well as environmental factors, complicating efforts to use to improve productivity. Previous research with spaced plants of tall fescue indicated a reasonably high realized heritability of 0.49 (Johnson and Li, 1999). Compared with the base population, populations selected for high led to reduced forage production, but populations selected for low did not have higher production. Jensen et al. (2004) found no correlation between forage production and in tall fescue cultivars at higher irrigation levels, but at low irrigation levels, was positively correlated with forage production. They concluded that selection for low (high TE) at low water levels would be likely to decrease rather than increase forage production. Thus, the relationship between forage production and in tall fescue is nebulous and can vary widely with environment.
The objectives of this research were to determine (i) if differences in high and low selections made on single plants would be maintained in solid seeded plots in the diverse environments of Ardmore OK, and Pullman, WA, and (ii) how related to forage production, leaf N, leaf C, and the C:N ratio among selected populations and a set of four tall fescue cultivars.
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
and other factors. The cultivars Alta (Hollowell, 1945) (PI 578712) and Fawn (Frakes and Cowan, 1974) (PI 578715) were used in previous studies of (Johnson and Li, 1999). ‘Jesup’ E+ (PI 592897), a recently developed cultivar with wild-type endophyte [Neotyphodium ceonophialum (Morgan-Jones & Gams.) Glenn, Bacon, & Hanlin comb. nov.] (Bouton et al., 1997), was included as a high-yielding check. The four populations of ‘Kentucky 31’ (K31) included a commercial source from Sharp Bros. Seed Co.1, Healy, KS (K31C), PI 561430 (K31 base population), PI 614890 (K31 high selection), and PI 614891(K31 low selection). The K31 base population was used for divergent selection of at Central Ferry, WA. The K31 base was collected from the Suiter farm in Menifee County, KY, considered the location for the first collection of what became Kentucky 31 tall fescue. Because endophyte viability of the K31 base population was lost in storage, it and the derived populations were selected without potential interactions with endophytes. After two cycles of divergent selection, the K31 high (PI 614890) and K31 low (PI 614891) populations were developed as outlined by Johnson and Li (1999). Briefly, leaves from 54 individual plants were sampled and analyzed for . Eight plants with the highest and lowest were selected. Six ramets from each selection were established in pots. The resulting 48 high and 48 low selections were placed in separate greenhouses, randomized, and allowed to intercross. Seed from each plant was harvested separately, and equal numbers of seed from each plant combined to form the seed of the first selection cycle (C1). Selection blocks of C1 high and low populations were then established in the field. As before, the eight plants with the highest and lowest mean values were selected from their respective populations, removed from the field, and crossed under greenhouse isolation to obtain seed of the second selection cycle (C2).
Plot Establishment
Plots were established at Pullman (46.72446 N and 117.13554 W) and Ardmore (34.19250 N and 97.08556 W). The seeding rate was 28.7 kg seed ha–1 at both locations. The Ardmore location was planted 13 Sept. 2001 in individual plots 4.6 by 1.5 m. The soil was a Heiden clay (fine, smectitic, thermic Udic Haplustert). The Pullman location was planted on 9 Apr. 2002. The plots in Pullman were 4.6 by 1.2 m, and the soil was Palouse silt loam (fine-silty, mixed, superactive, mesic, Pachic Ultic Haploxeroll). The seven entries were randomized in complete blocks with three replications at both locations, and plots were maintained under dry-land conditions according to locally recommended fertility rates and procedures.
Data Collection
At Ardmore, plots were harvested in January 2003, May 2003, January 2004, July 2004, December 2004, and June 2005. At Pullman, harvests were in September 2002 and July 2003, 2004, and 2005. For each date, forage from each plot was cut about 7.5 cm above ground level, removed, and dried; forage production was calculated on a dry weight basis.
Before each harvest, samples of upper, fully emerged leaves were collected from 10 to 12 plants per plot, dried at 70°C to constant weight, and ground to pass through a 0.5-mm screen. Carbon isotope discrimination, percentage C, and percentage N were determined at the Augustana College (Sioux Falls, SD) stable isotope laboratory. The finely ground leaf samples (2.5–3.5 µg) were weighed and, with standards of known composition for 13C, percentage N, and percentage C, placed in an autosampler and analyzed as outlined by Read et al. (1991). Stable carbon isotopes, 
1313C (the ratio of 13C/12C relative to the PeeDee belemnite standard), were obtained. Values of 13C were converted to carbon isotope discrimination () using a 13C for air of –8 per mil as described by Farquhar et al. (1989).
Plant stands were monitored as outlined by Hopkins et al. (1993). At Ardmore this was done post-harvest in spring 2002 and 2004, winter 2005, and spring 2005. Briefly, a 1-m2 grid divided into 25-cm quadrants was placed randomly over plots. The number of quadrants without live plants was counted and used to calculate percentage stand. This was repeated twice for each plot. At Pullman stand data were similarly taken in the fall of 2002, 2003, and 2005.
Periodic estimates of endophyte infection in entries were completed on all plots. Sampling dates for Ardmore were October 2002 (4 plants plot–1), April 2005 (5 plants plot–1), and July 2005 (6 plants plot–1) for each of the three blocks, resulting in a total of 45 samples for each entry. At Pullman, 6 plants plot–1 were sampled in June 2002, June 2004, and 20 July 2005 for each block, resulting in 54 plants sampled per entry. On a single culm from different plants, the presence or absence of endophyte was determined using a commercial immunoblot tiller test kit (Hiatt et al., 1999).
Data Analysis
The experiment at each location was randomized in complete blocks with repeated measures. The repeated measurements were leaf , percentage leaf C, percentage leaf N, the C:N ratio, plot stands, and forage production. Thus, for a given repeated measurement, such as forage production, the variation was partitioned into blocks, entries, the block x entry interaction, harvest date, and the entry x harvest date interaction. The error term for testing entry effects was the block x entry interaction. The residual error was used for testing the harvest date effect and the harvest date x entry interaction. Data from the Ardmore and Pullman locations were analyzed separately using general linear models (SAS Institute, 2003), assuming fixed effects for all factors except blocks, which are random. Simple linear correlation analysis, and partial correlation, controlling for harvest effects, was also completed on mean values at each harvest. Treatment differences were declared significant at P < 0.05.
|
RESULTS AND DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
The contrast in climate between Ardmore and Pullman is striking (Table 1 ). Long-term data show annual average, maximum, and minimum temperatures from 7 to 9°C higher for Ardmore than Pullman. The average maximum temperature in the period from July to September in Ardmore was always higher than 30°C during the experimental period (data not shown), and the long-term average was 33°C. For the same period, the average maximum at Pullman was 26°C (data not shown), and the long-term maximum temperature was 27°C (Table 1). Temperatures near 25°C are considered optimal for tall fescue growth, with temperature of 30°C and higher detrimental (Robson, 1972). Thus, high summer temperatures commonly experienced in Ardmore, and for the Southern Plains in general, are well above the optimum for tall fescue growth.
|
During the experimental period, Pullman was dryer than average for all years except 2003, when total precipitation was 54.3 cm. For 2002 and 2004, Pullman had 40.8 cm of total precipitation each year compared with the average, 53.8 cm. For Ardmore, precipitation was 91.6 cm in 2002 and 105 cm in 2004 but sharply lower in 2003 at 64.5 cm.
Analyses of Variance
Average forage production per harvest date was more than three times higher at Pullman than Ardmore (Table 2
). However, there were six harvest dates at Ardmore and only four at Pullman. Total production over the 4-yr experimental period was 12.1 Mg ha–1 for Ardmore and 26.8 Mg ha–1 for Pullman. Perhaps the most important factor in the lower forage production at Ardmore was the high summer temperatures discussed above. Even though there was some stand decline at Pullman compared with Ardmore, Pullman had higher forage production.
|
effects at both locations. Among all traits, the only significant harvest date x entry interaction was for forage production at Ardmore.
Harvest Date Effects
At Ardmore forage production at the various harvests was highly variable. The highest forage production at Ardmore was for the January 2003 harvest, and the lowest was for the December 2004 harvest (Table 3
). The low production in December 2004 suggests limited regrowth after the July 2004 harvest. Although leaf N, C, and the C:N ratio differed among harvests at Ardmore, dry weight variation was not clearly associated with any of those factors.
|
values (lower TE) across harvests (Table 3). This suggested a link between higher forage production and lower TE associated with different Pullman harvests. The high production in 2003 had high N, but a link between N and production was not consistent across harvests. As at Ardmore, there was no apparent association with C or the C:N ratio with forage production at different harvest dates at Pullman.
Entry Effects
The presence of Neotyphodium endophytes in entries revealed infection rates at Pullman generally consistent with expectations; the selections had no infection to very low infection, and Jesup E+ had a high infection rate (Table 4
). The entry K31C had a relatively low infection rate at both Pullman and Ardmore. The infection rates of 20 and 29% for the K31 low and high selections at Ardmore were not expected. Previous work (Johnson and Li, 1999) and endophyte data at Pullman (Table 4) indicated that these had either very low infection rates or were infection free. Thus, there must have been a source of plant contamination in the Ardmore plots, perhaps associated with high-intensity rains that occurred shortly after planting that may have transported seeds during establishment. However, endophyte infection levels did not change over time at Ardmore or Pullman. Although endophyte infection rates at Ardmore differed from those at Pullman for the K31 low and high entries, there was no obvious association between those infection rates, selection, and forage production.
|
entry had higher forage production than the K31 high entry, as observed at Pullman (Table 5
). As expected, Jesup E+, a high-yielding, recently developed cultivar, was among the entries with generally high forage production. The K31 low entry also had relatively high yield, especially at Pullman. On average, the K31 high entry ranked lowest in forage production (Table 5). Johnson and Li (1999) found in spaced plants that Alta had greater forage production than Fawn even though they had similar values. In the current study, Alta had lower than Fawn at Pullman but forage production did not differ. At Ardmore, for Alta and Fawn did not differ, but forage production was higher for Alta in January 2003 at Ardmore (data not shown). Thus, there was no consistent association between and forage production between Alta and Fawn.
|
selection, base population, and K31 high selection, made on spaced plants in eastern Washington, had low, intermediate, and high at both sites in solid seeded plots (Table 5). This is consistent with the conclusion by Johnson and Li (1999) on spaced plants that is a heritable trait in tall fescue. It also shows that the plant-endophyte contamination at Ardmore (Table 4) had no apparent effect on the rankings. At Ardmore neither N nor C differed among entries, but there were differences in the C:N ratio. Nevertheless, the relationship between C:N ratio and forage production was nebulous.
For forage production at Pullman, the K31 low and high populations differed and the K31 base was intermediate (Table 5). This pattern was consistent with low promoting forage production. At Pullman, production averaged 9.6% higher for the K31 low selection than the K31 high selection, but differences between the selections and the base population were not significant at P < 0.05 (Table 5). At Pullman, neither leaf N nor the C:N ratio differed, but there were differences in leaf C (Table 5). For example, Alta and the K31 low selection both had high yields, but Alta had high leaf C and K31 low selection had low leaf C. Thus, differences in C were not generally associated with production.
Correlation
Linear correlation over harvest dates at Ardmore showed that was negatively correlated with forage production (r = –0.32, P < 0.05, n = 42). This suggests that lower , and presumably higher TE, contributed to higher forage production. At Pullman the correlation between and forage production was strongly positive (r = 0.73, P < 0.01, n = 28), suggesting that higher and therefore lower TE contributed to higher yield. These contradictory results show the influence of location and harvest date on the and forage production relationship. The positive correlation between and forage production at Pullman suggests that in high production years, such as 2003, stomatal conductance and internal leaf photosynthetic capacity are high, resulting in generally enhanced growth. Under these conditions, the internal leaf CO2 concentration would be higher. resulting in higher and TE (Johnson and Li, 1999).
Others have observed positive correlations between and production (Condon et al., 1987; Araus et al., 2003) when water was not limiting or under relatively wet conditions. Jensen et al. (2004) found a positive correlation between and production at low water levels in tall fescue. However, they also showed that as water application increased, there was an overall increase in . Under dryer conditions, differences in plant water status among entries could develop. Plants with a higher water status would be expected to have higher ; a positive correlation between and production could result in association with drought avoidance mechanisms, such as a deeper or a more efficient root system. If plant water status among genotypes varied with such mechanisms, important genetic differences in may be obscured. This is the risk of selection for under severe drought. Moreover, less-severe stress associated with atmospheric water deficits and milder, yet significant, drought periods are more common than severe drought and likely to have application to a wider set of conditions.
Partial correlations between and forage production at both Ardmore and Pullman, controlling for harvest date effects, were not significant, suggesting that environmental differences from harvest to harvest within the locations were responsible for the correlations between and forage production that were observed across harvests. Even with the strong environmental influences of both location and harvest date, the genetic differences in among selections were always maintained (Table 5). Thus, differences in high or low populations selected from single plants should be maintained in solid seeds stands and in diverse environments.
Using only the low and high selections and the base population, partial correlations of with forage production at Ardmore were not significant, but at Pullman they were significant (r = –0.59, P < 0.05, n = 12). Since selections for were made in eastern Washington, perhaps they were generally less well adapted to Oklahoma, making a forage production to relationship less likely. Nevertheless, selection for low was never detrimental to forage production at either Pullman or Ardmore. The low selection had higher production than the high selection at Pullman, and for two of the six harvest dates at Ardmore. So it appeared that there was some potential for higher forage production with low selection, and that higher TE would not lead to reduced forage production.
Given that low may increase forage production only modestly and only in some environments, it probably should not be the main focus of a breeding program despite its heritable nature. Experience with variation in outcrossing grasses (Johnson and Bassett, 1991; Johnson and Li, 1999) suggests that selection for low to obtain higher TE within advanced tall fescue lines is possible. Johnson and Li (1999) found that forage production of spaced plants did not increase and may even decrease between cycle 1 and cycle 2 selections for low . Thus, one cycle of selection would likely be most efficient and would also be relatively inexpensive. The data also suggest that this could be done on spaced plants, with the expectation of similar in solid seeds stands.
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
Received for publication December 4, 2007.
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Vadose Zone Journal | |||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
